I authored the recent change you're commenting on.  I agree with your
points; my only defense is consistency.  Several other nearby
assertions used the older try-catch based setup.

I'll fix the spot you objected to, and file a JIRA around cleaning
this up more broadly.  Having this elsewhere in the code encourages it
to creep in more.

Best,

Jason
On Fri, Jul 6, 2018 at 12:58 PM Chris Hostetter
<hossman_luc...@fucit.org> wrote:
>
>
> these tests should really be using...
>
>   SolrException e = expectThrows(() -> {...});
>
> ...and ideally we should be making assertions about the exception message
> as well (ie: does it say what we expect it to say? does it give the user
> the context of the failure -- ie: containing the "non_numeric_value" so
> they know what they did wrong?
>
>
> :    private void validateCommonQueryParameters() throws Exception {
> :      ignoreException("parameter cannot be negative");
> : +
> : +    try {
> : +      SolrQuery query = new SolrQuery();
> : +      query.setParam("start", "non_numeric_value").setQuery("*");
> : +      QueryResponse resp = query(query);
> : +      fail("Expected the last query to fail, but got response: " + resp);
> : +    } catch (SolrException e) {
> : +      assertEquals(ErrorCode.BAD_REQUEST.code, e.code());
> : +    }
> : +
> :      try {
> :        SolrQuery query = new SolrQuery();
> :        query.setStart(-1).setQuery("*");
> : @@ -1228,6 +1238,15 @@ public class TestDistributedSearch extends 
> BaseDistributedSearchTestCase {
> :      } catch (SolrException e) {
> :        assertEquals(ErrorCode.BAD_REQUEST.code, e.code());
> :      }
> : +
> : +    try {
> : +      SolrQuery query = new SolrQuery();
> : +      query.setParam("rows", "non_numeric_value").setQuery("*");
> : +      QueryResponse resp = query(query);
> : +      fail("Expected the last query to fail, but got response: " + resp);
> : +    } catch (SolrException e) {
> : +      assertEquals(ErrorCode.BAD_REQUEST.code, e.code());
> : +    }
> :      resetExceptionIgnores();
> :    }
> :  }
> :
> :
>
> -Hoss
> http://www.lucidworks.com/
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
> For additional commands, e-mail: dev-h...@lucene.apache.org
>

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org

Reply via email to